E2E Deployment Test Failure: Aspire.Deployment.EndToEnd.Tests.AcaStarterDeploymentTests.DeployStarterTemplateToAzureContainerApps

## Failing Test

`Aspire.Deployment.EndToEnd.Tests.AcaStarterDeploymentTests.DeployStarterTemplateToAzureContainerApps`

## Workflow Run

[Deployment E2E Tests — 2026-06-08 (run 27116191858)](https://github.com/microsoft/aspire/actions/runs/27116191858/job/80024042070)

## Symptom

```
Xunit.MicrosoftTestingPlatform.XunitException: System.InvalidOperationException :
Pipeline failed unexpectedly. Terminal output:
'Invalid value: "infraacrzoo7q3iyi3v7m.azurecr.io/webfrontend:aspire-deploy-20260608044239":
Get "https://infraacrzoo7q3iyi3v7m.azurecr.io/v2/": dial tcp: ...
04:47:37 (pipeline-execution) ✗ [ERR] Deployment failed: Error code = ContainerAppOperationError,
Message = Failed to provision revision for container app 'webfrontend'.
✗ provision-webfrontend-containerapp — Failed to provision webfrontend-containerapp
```

## Root Cause

The ACA control plane validates `template.containers.webfrontend.image` by doing a live DNS lookup of the registry hostname. When `aspire deploy` provisions a fresh Azure Container Registry and then immediately provisions container apps that reference images in that registry, the new `.azurecr.io` record has not yet propagated to ACA's DNS resolvers, producing `no such host`.

This is the same DNS-propagation race tracked in #17739 (`AzureStorageDeploymentTests.DeployStarterWithSharedAzureStorageReferences`) and the analogous FrontDoor failure filed in this batch.

## Notes

- Infrastructure / timing-side failure; needs a retry or readiness probe on the registry hostname inside the ACA provisioning step.
- Auto-filed by the deployment-test analysis workflow.